Author: Fivethousandoverlibor

(This is a guest post from an anonymous professional natural gas trader.)

The EIA released an as-expected decline in US production, amounting to an average of -0.40 Bcf per day for the Lower 48 States series for the period Jan 2009 through Jan 2010.  If this was an underwhelming figure, the fact that the report now shows the latest month estimated, February, to be the new peak in US production, at 63.9 bcf/day, must have been frustrating.  The 1 bcf/day increase over January does support the idea presented here that a more responsive supply-estimation method would not manifest bullishly given the record increase in Baker Hughes’ gas rig count.

Yesterday was instructive for those trying to understand production data’s place in the hierarchy of natural-gas fundamental data, for if the 914 revisions were uninspiring, the weekly storage report was
anything but.  The build of 83 Bcf for gas-week ending Apr 23 signals a quickly loosening supply/demand balance and emphasizes the importance of storage data over lagged production data that still do
not fit well with demand assumptions and the separately-assembled storage data.  While the production news was bearish, it took a back seat to the 297 Bcf we’ve built since we began building inventories 6
weeks ago.  This compares to a build of 156 Bcf over the same period last year – a record then.  So, not only have we beaten last year’s record start to the injection season, we’ve doubled it.  There is no
doubt that the weather since mid March has been very mild and greatly contributed to this storage result, but an aberrant spring is not akin to an aberrant winter or summer when talking energy demand.  The
demand sensitivity of gas to temperature swings is not linear and from 50F to 75F is fractions of the sensitivity you see from 25F to 50F and from 75F to 100F.  On a weather-adjusted basis, yesterday’s report was actually bearish against last year’s build of 82 bcf for the same week.  This is significant, because the time period one year ago in gas was thought to be the hallmark of poor supply/demand dynamics.

The supply/demand balance, as understood through storage, appears to be poor and appears to be worsening (this is particularly poor news for those carrying gas through time in UNG, as contango will expand in a loosening environment).  It’s prospects into May do not look bright given that we rolled the May contract this week at $4.27 – 95 cents (30%) above last year’s May expiry and 43 cents above April.  This is a level that will offer pause for many of those executing economic options between coal and natural gas for power generation and encourage any existing price-sensitive marginal production.  Price responsiveness – on both sides of the equation – has become a defining characteristic of gas in the new shale era.  May at $4.27 will likely not help to tighten the balance, but should help to highlight the importance of storage data over production data in assessing the US gas market.

Join the conversation about this story »

(This guest post comes courtesy of a pseudonymous natural gas trader)

The United States Gas Economy is simple: Supply-Demand=Storage.  Every day of the year.

The EIA releases an aggregated survey-based estimate of US natural gas storage weekly – produced from submissions of form EIA-912 by storage and pipeline operators.  This is an original data source from which our understanding of current US storage, and therefore the supply/demand balance, is predicated.  This is why this weekly number garners the attention that it does: it is, in large measure, the genesis of nearly any understanding of the US Natural Gas market.  Much has been made of the burgeoning “balancing item” in EIA data; a favorite of the “gas is in decline, damn it” crowd.  The “balancing item” is a plug figure enlisted by the EIA to balance their supply estimates with their demand estimates as they then both relate to…weekly storage data.  In other words, a popular argument is EIA 914 is materially overstating production because we would actually have significantly higher storage inventories if EIA 914 estimates were correct.

Storage data is not in question.  Is it free from errors?  Not a chance.  Are they huge?  Could be.  I am not warrantying them in any manner.  But, I am not aware of any current challenge to storage data and it is not in any way being questioned by Thursday’s release of revisions to EIA 914 data.  This is important.

If storage data is not in question, meaning, there is no challenge that current NG inventories sit at 1,829 Bcf, then whatever revisions are made to production data in EIA-914 will necessarily result in commensurate downward revisions to demand estimates.  It does not, in any way, alter our understanding of the supply/demand balance, it simply shifts the legs that constitute the balance.  Any shift we find on Thursday will be important.  If we see a decline in production, principal deductions will be that production was more sensitive to the largest abandonment of gas rigs in the history of mankind and that demand was more sensitive to the greatest liquidity-driven recession in the history of mankind.  This will not require an extended walk with reasoning should it be called upon.  Some would argue it makes for a more realistic script.

Back to the 1,829 Bcf and April 16, 2010.  This proves to be the highest level of NG inventories for this time of the year; it is higher than last year by roughly 95 Bcf.   This is interesting to some, as we did see a record draw for Dec 1 –  Feb 28 this past winter – largely driven (this is becoming better understood with each storage report this spring) by marvelous winter events (research the Arctic Oscillation: you’ll find grown scientists insisting upon a -4.5 sigma event this past winter).  DFW got a foot of snow in mid Feb.  Washington’s Dulles got 42 inches of snow, the entire accumulation from the previous 4 years, in 6 February days.  The 2.1 Tcf pull from storage Dec 1 – Feb 28 was the subject of considerable conversation as it became forecasted by mid Feb and realized into early March.  How much was due to not only the severity of the cold, but the location of the cold?  Shifts in demand are not uniform across the country for a given drop in temperature.  Answers not found through some healthy mixture of hard-nosed analysis and incantation would be found through time: spring would deliver some clarity.

March registered a net draw of 45 Bcf.  This is followed by a current estimate for a net build between 290 and 305 Bcf for April.  March is a record-low draw; April will prove to be a record-high build.  While temperatures were very mild (both March and April looking to be the 2nd warmest over the past 5 years), when adjusting for weather, given the storage result we have for the last 5 weeks it becomes difficult to husband the storage results from the winter (through Apr 16 – a build of 214 Bcf, the largest on record, beating last year’s record result for the same period (days 71 – 106 of the year) of a build of 79 Bcf).   Further, it strongly suggests that winter storage results were not driven by the material tightening of the supply/demand balance but rather the idiosyncratic matrix of temperature and place this past winter.

This is not a call on price.  There is healthy doubt swirling around the idea that shale gas has changed exploration and production into a manufacturing process.  But the key element that directly influences price – the supply/demand dynamic, a.k.a. storage – will not be under review on Thursday with the release of revisions to EIA 914.   We are likely to simply learn that production and demand were reasonably more responsive to the historical elements that have driven them since Jan 2009.  Given the fact that until last week we hadn’t seen a decline in gas rigs since Christmas, mounting the largest net add to gas rigs over a four-month period in Baker Hughes gas-rig-data history, the potential for a more responsive production data set to changes in rig counts doesn’t immediately wax bullishly.

Join the conversation about this story »

(This guest contribution comes from an anonymous veteran natural gas trader. It concerns the rot of the infamous US Natural Gas fund ETF “UNG.”)

While I cannot claim that every source of financial media has failed in attempting to fully assess the rot that represses UNG (UNG), I can assert that every source I have been made aware of has.  Although most are not incorrect, they are incomplete, often going 2 of the required 200 feet to fully understand what is wrong with UNG and every other ETF that employs a long-only, prompt-only consumable-commodity futures strategy.

UNG carries prompt-month gas futures, perpetually, through time.  Carry and carry and carrying away, in essence morphing long-term UNG holders into Natural-Gas Sherpas.  While none of UNG’s activities are in the physical arena, the profit/loss profile of their strategy is identical to holding physical length for 30 days and then contracting to carry it – at going market rates – for another 30 days.  Rinse and repeat.

Carrying natural gas (and most other consumable commodities) is not an activity free of expenses.  Storing this fuel costs money.  How much?  It depends upon who you ask and what modified form of “carrying” you are involved in.  If you owned a parcel of physical gas and contracted to hold it in a storage facility until January, you would pay an explicit, determined rate to an actual pipeline/storage operator.

If, however, you are involved in a strategy of constantly holding prompt Natural Gas futures, your storage cost (you probably didn’t even know you had one) is far less obvious.  This is where that frightfully odd sounding word “contango” comes in – a property absolutely foreign to almost all non-futures participants until the advent of these Bouncing Betty ETFs.  

Contango refers to a specific shape of a forward futures curve – where deferred deliveries are more dear than those previous.  Contango exists for many reasons but primarily to encourage storage of a consumable commodity.  Absent this storage we would all suffer perpetual lapses of surplus and deficit of supply in the given commodity’s spot market and the concomitant extreme price volatilities that would result (watch spot electricity prices for a vision of a future without storage).  For many consumable commodities, contango is the norm.  Natural gas is no different: over the past 5 years the spread between the first and the second months has had a contango settlement 1,178 of the 1,257 trading days – or 94% of the time.  Since NG futures inception at NYMEX in April 1990, 80% of all daily settlements have been in contango.

UNG, therefore, (and any other long-only, prompt-only consumable-commodity futures fund) has twin exposures:
1)      The obvious: Long Natural Gas – in the form of the futures, swaps, and total return swaps it carries
2)      The not-so obvious: Short Natural Gas Storage – in the commitment it has to sell prompt and buy next-to-prompt gas each month.

It’s the short-natural-gas-storage exposure (wholly naked at that) that many completely missed when this ETF first rolled out and that many in the media now inaccurately describe with statements such as “UNG bought fantastically rich forward contracts, and then rode them all the way down to spot”.  This misses the point – the problem with UNG’s structure is virtually independent of the price she secured her previously-purchased gas.   The problem is with a vastly-dominating positive prompt 30-day storage cost.  To wit: the largest singular monthly problem that UNG ever suffered was taking her OCT 2009 gas 30-days forward into NOV 2009 last year, and “paying” $1.00 per MMBtu to do it.  On $3.20 gas (OCT 2009 avg price on their roll back in September), that’s a 31% notional carry cost – for 30 days.  They bought that OCT at around $3.75 a month earlier – a price near 7-yr lows at the time.  Fantastically rich is not a prerequisite for the pain of UNG.

Storage costs for UNG will always accumulate faster than the gas that UNG carries can appreciate over time.  Why?  Because natural gas prices oscillate about a marginal cost to produce and that marginal cost to produce does not advance ad infinitum. Things, like technology, interrupt advancing costs (horizontal drilling being a strong example). Storage costs, on the other hand, will steadily accumulate throughout time with little interruption. As stated previously, the front 2-month spread in NG has been in contango 80% of the days since inception in 1990.  This worsens to 87% and 94% for the past 10 and 5 years, respectively.  This eventually renders UNG worthless and is expressly why no entity/fund/person in the futures market has ever deployed a perma-long, perma-prompt strategy prior to the advent of these ETFs.

Now, UNG never explicitly “pays” these “storage costs”, they simply lose gas each roll – they pay for the storage by taking less gas with them every 30 days.  But the exposure is short storage, as that is what UNG is, a vehicle that carries gas for 30 day clips that will continue until the fund eventually “spends” all of its capital on NG storage.  This will manifest with UNG simply running out of gas – literally.

 UNG works well for those seeking short periods of prompt natural-gas exposure.  It is utterly ruinous for long-term holders – as long-term length in UNG is nothing but being a Natural-Gas Sherpa.  UNG is capable of sustaining advancements in price, in so far as NG prices are advancing faster than storage costs accumulate.  But in this new era of Shale Gas, where storage is the constrained variable, not production, the prospects for such developments are bleak.    With UNG in the Shale Gas Era, you are short the constrained and long the plentiful – not an advantageous place to be.

Join the conversation about this story »